Proteaphagy is specifically regulated and requires factors dispensable for general autophagy

Abstract

Changing physiological conditions can increase the need for protein degradative capacity in eukaryotic cells. Both the ubiquitin-proteasome system and autophagy contribute to protein degradation. However, these processes can be differently regulated depending on the physiological conditions. Strikingly, proteasomes themselves can be a substrate for autophagy. The signals and molecular mechanisms that govern proteasome autophagy (proteaphagy) are only partly understood. Here, we used immunoblots, native gel analyses, and fluorescent microscopy to understand the regulation of proteaphagy in response to genetic and small molecule-induced perturbations. Our data indicate that chemical inhibition of the master nutrient sensor TORC1 (inhibition of which induces general autophagy) with rapamycin induces a bi-phasic response where proteasome levels are upregulated after an autophagy-dependent reduction. Surprisingly, several conditions that result in inhibited TORC1, such as caffeinine treatment or nitrogen starvation, only induced proteaphagy (i.e., without any proteasome upregulation), suggesting a convergence of signals upstream of proteaphagy under different physiological conditions. Indeed, we found that several conditions that activated general autophagy did not induce proteaphagy, further distinguishing proteaphagy from general autophagy. Consistent with this, we show that Atg11, a selective autophagy receptor, as well as the MAP kinases Mpk1, Mkk1, and Mkk2 all play a role in autophagy of proteasomes, although they are dispensable for general autophagy. Taken together, our data provide new insights into the molecular regulation of proteaphagy by demonstrating that degradation of proteasome complexes is specifically regulated under different autophagy-inducing conditions.

Keywords: autophagy; proteaphagy; proteasome; proteasome inhibitor; protein degradation; starvation; target of rapamycin (TOR); vacuole; yeast.

Figure 1

Rapamycin induces a bi-phasic proteasome response.A, yeast expressing α1-GFP were grown in YPD medium and treated with 200 nM rapamycin. Equal volumes of lysates were separated on native gels and imaged for GFP (top panel) or peptidase activity in the presence of 0.02% SDS using the proteasome substrate suc-LLVY-AMC (lower panel). The graphs show the quantifications of the levels of GFP and LLVY peptidase activity corresponding to the RP₂-CP species on native gel, both normalized using Pgk1 signal from three experimental replicates. The error bars represent SEM. p values for upregulation at the half hour timepoint were 0.006 for GFP and 0.03 for LLVY-AMC using unpaired t test. B, the samples treated as in (A) were denatured and separated on SDS-PAGE and immunoblotted for GFP. The upper band shows α1-GFP and lower band is “free” GFP resulting from α1-GFP cleavage by vacuolar hydrolysis. Pgk1 was used as a loading control. The data presented are representative of three independent experiments. C, WT and atg7Δ yeast expressing α1-GFP or Rpn1-GFP were inoculated in YPD medium at A₆₀₀ of 0.5 and grown to log phase (∼4 h). The cells were treated with rapamycin as in (A) and lysed at indicated time points. α1-GFP, Rpn1-GFP, and cleaved free GFP (indicative of vacuolar targeting) were monitored by immunoblotting for GFP. Pgk1 was used as a loading control. The graph shows quantification of the immunoblots from 2 independent experiments and the error bars represent SEM (D) Rpn1-GFP expressing yeast was stained for the vacuole membrane using FM4-64 and the nucleus using Hoechst 33,342. Microscopy was performed at log phase and following 24 h growth in rich media (YPD) (top). Rpn1-GFP expressing yeast growing logarithmically were incubated with the vacuole lumen marker CMAC-arg (bottom). The scale bars represent 0.5 μm and data presented are representative of consistently observed staining patterns from many independent experiments (E) microscopic analysis of yeast collected at log phase and 24 h after rapamycin treatment. In top image, the arrowheads point to nuclei and arrows to vacuoles. In rapamycin-treated cells, the filled arrowheads indicate cells with vacuolar fluorescent signal, whereas open arrowheads show subset of nonresponding cells. The scale bars represent 0.5 μm. The values indicate the percentage of cells in WT and atg7Δ with vacuolar GFP signal after rapamycin treatment. Vacuolar GFP signal was rarely (<1 %) observed in nontreated cells. Approximately, 55% of WT cells showed vacuolar localization of proteasomes after rapamycin treatment (n > 100, SEM= 1.3 from three independent experiments). CP, core particle; RP, regulatory particle; YPD, yeast-peptone dextrose.

Figure 2

Tunicamycin induces a bi-phasic proteasome response.A, yeast expressing α1-GFP were treated with tunicamycin (6 μM) for indicated time periods. The cells were collected, lysed under native conditions, and equal amounts of lysates were loaded on native gel. After separation, gels were imaged for GFP fluorescence (top) or suc-LLVY-AMC peptidase activity in the presence of 0.02% SDS (bottom). Quantifications show the amount of GFP signal and LLVY peptidase activity associated with the RP₂-CP proteasome complexes. GFP and LLVY activity were normalized to Pgk1 intensity using SDS-PAGE immunoblots of the same samples. Three independent experiments were used and the error bars represent SEM. B, WT and atg7Δ yeast were treated with tunicamycin and the lysates analyzed as in (A) (left) and denatured for Western blotting (right). The data presented are representative of three independent experiments (C) WT, atg39Δ, atg40Δ, and atg39Δ atg40Δ yeast expressing Rpn1-GFP were treated with tunicamycin as in (A). Two absorbances of cells were lysed using the alkaline lysis method. The samples were separated on SDS-PAGE and immunoblotting for GFP and Pgk1 was carried out as described above. The blots are representative of three independent experiments. CP, core particle; RP, regulatory particle.

Figure 3

Proteaphagy is distinct from general autophagy.A, yeast strains expressing GFP-atg8, GFP^OE, or α1-GFP, were grown in YPD and subsequently starved 24 h for nitrogen (-N), amino acids (-AA), or phosphate (-PO₄). The equivalent of 50 absorbances of cells was lysed by cryogrinding. Equal volumes of the lysate were blotted for GFP and Pgk1. The value below immunoblots indicate the free GFP signal as percentage of free + fused GFP. The blots presented are representative of three independent experiments (B) localization of fluorescent proteins in starved cells from (A) was monitored by microscopy. The scale bar represents 5 μm and data presented are representative of three independent experiments. C, WT and atg11Δ yeast were grown to log phase in rich medium and switched to phosphate-starvation medium. Two absorbances of cells were harvested at indicated timepoints and analyzed as above. The data presented are representative of three independent experiments. YPD, yeast-peptone dextrose.

Figure 4

Proteaphagy requires factors dispensable for general autophagy.A, WT, atg11Δ, atg17Δ, atg11Δ atg17Δ yeast expressing Rpn1-GFP were starved for nitrogen or treated with rapamycin for 24 h. The data presented are representative of three independent experiments. B, WT and mpk1Δ cells expressing Rpn1-GFP or α1-GFP were starved for nitrogen and 2 absorbances were harvested at indicated time points. Immunoblotting for GFP and Pgk1 were performed, as described above. We observed an ∼35% reduction in cleaved GFP from Rpn1 and ∼33% from α1 (p value 0.067 and 0.0005, respectively from unpaired t test) in the mutants compared with WT 24 h after starvation. The data presented are representative of three independent experiments. C, WT, mkk1Δ, mkk2Δ, and mkk1Δ mkk2Δ cells expressing Rpn1-GFP were starved for nitrogen and 2 absorbances of cells were harvested at 24 h. Immunoblotting for GFP and Pgk1 were performed, as described above. The data presented are representative of three independent experiments. D, WT and mkk1Δ mkk2Δ cells expressing Rpn1-GFP or α1-GFP were starved for nitrogen and 2 absorbances of cells were harvested at indicated time points. From three independent experiments, we observed an ∼39% reduction in cleaved GFP from Rpn1 and an ∼40% reduction from α1 (p = 0.030 and 0.119 respectively from unpaired t test).